By using a generalized inversion technique and the finite-element method, permanent changes in displacement caused by the Alaska earthquake of 1964 are analyzed to investigate a regional variation in the faulting mechanism of the extensive fracture system.
In section 2, the fracture system is modelled by a multiple-fault system composed of four rectangular faults in an elastic half-space. The results of the inversion analysis for the multiple-fault system show that the Alaska earthquake consists of a main faulting of a low-angle underthrust and a subsidiary surface faulting of a somewhat high-angle underthrust with a little left-lateral component. The main fault extends from Kodiak Island through Prince William Sound to Kayak Island with a total length of 600 km, and the dip-angle (δ) and dislocation (D) in the corresponding regions are δ=25°NW and D=19m, δ=7°NW and D=10m, and δ=20°NE and D=8m respectively. For the subsidiary fault, the optimal estimates are δ=31°NW and D=12m. The strike direction (N36°E) is parallel to that of the main fault in the Kodiak Island region.
In section 3, setting a vertical section to be perpendicular to the strike direction in the Kodiak Island region, the fracture system is treated in a framework of a plane-strain, finite-element approximation, where each fault surface is simulated by a sequence of dislocated double nodes. From the inversion analysis for the two-dimensional fault model, it was found that the dislocation along the main fault surface has a broad peak of 30m at a depth of 20km and decreases monotonously up to a depth of 60km. The profile in the shallower part is obscure, because of the lack of the data in the Gulf of Alaska. For the subsidiary fault, the dislocation has a peak of 12m close to the earth's surface and decreases steeply at a depth of 10 km. The profiles of the observed displacement fields across the fracture system are well interpreted by these dislocation functions varying with depth.

View full abstract

On the basis of the elastic dislocation theory, formulas are derived for the displacement, strain and stress fields of an arbitrarily situated, two-dimensional strike-slip fault in a laterally inhomogeneous medium. It is assumed that the elastic half-space is made up of n+1 isotropic and homogeneous media being in welded contact and separated by n vertical interfaces. The explicit expressions of coefficients which appear in the formulas are listed for the case n=2. They are checked with the previous results obtained with the help of the method of images. As an application of the formulas derived, the surface displacement field of a vertical strike-slip fault is determined. Three different models are analyzed on the basis of the data concerning medium structure along a section of the San-Andreas fault near Bear Valley. In the presence of lateral inhomogeneities of the medium, it is shown that the elastic field of an earthquake should exhibit distinct asymmetry. This conclusion is confirmed by the existing geodetic data. In particular, the Tango earthquake of 1927 is discussed in greater detail.

View full abstract

The deep crustal structure is studied by using waves converted at the Mohorovicic and Conrad discontinuities. The data are earthquakes recorded by a short period seismometer with high sensitivity for frequencies ranging from 1 to 200Hz for a period of about three months from 1974 to 1975 at Tsukuba Seismological Observatory of the Earthquake Research Institute. The S-P times are about 5 to 10 seconds. The magnitudes are 1.2 to 3.4. The epicenters are located in southwestern Ibaraki, off the coast of northern Ibaraki, near the Tochigi-Ibaraki border and around Chiba city. The focal depths are about 45 to 79km.
Crustal thickness is estimated to be about 23 to 26km at four sites around Tsukuba. A Conrad discontinuity is also found at two sites in the same district. Its depth is about 9km and 20km, respectively. It is shown that the converted waves are a very useful tool for determining the deep crustal structure.

View full abstract

A static dislocation model of the Sanriku earthquake of 1933 is proposed which is consistent with the initial motion of the tsunami. The wave equations of a tsunami were solved numerically on the displacement fields for various fault parameters and the computed time histories of the water surface elevation at the coast were compared with the observed initial motions. The best fit parameters were determined by trial and error. It is a normal fault reaching the earth's surface and having a dip of 30° toward the west, a dimension of 270km (length)×70km (width) and a dislocation of 2-5m. The assumed rigidity of 7×10¹¹ dyne/cmcm2 gives a moment of (3-7)×10²⁸ dyne-cm. The fault of this model has a longer, narrower fault surface and smaller dip angle than that of the seismic model (KANAMORI, 1971a). The dip angle of the present model is in fairly good agreement with that of the deep-focus seismic plane in this region. It is understood that the upper part of the sinking slab slid down with the earthquake and it contributed to make the oceanic plate slide down as a whole under the island arc. Thus it is shown that the tsunami data is useful in making a fault model.

View full abstract

A systematic variation in the mode of seismicity is presented as a process accompanying earthquakes of magnitude M ranging from 4.7 to 5.2 occurring periodically with a recurrence time of 8-15 years in the vicinity of Wakayama city, northwestern part of the Kii Peninsula, Central Japan. Earthquakes of magnitude M larger than 4.5 can be reasonably classified into the representative major events in this area where few shallow earthquakes are of magnitude M larger than 5.0.
By a continuous micro-earthquake observation undertaken in the area, a distinct aseismic zone, a so-called seismicity gap, as well as a noticeable lineament of epicentral distributions across the seismicity gap have been confirmed to exist in the area. The seismicity gap, occupying an area of about 50km², appears to be evident 2-3 years before a major earthquake occurrence showing a remarkable contrast to the surrounding area characterized by earthquake swarms. The occurrence of a precursory earthquake of magnitude M as large as 4.3 at the southern border of the seismicity gap is identified as a reliable indication of a subsequent major earthquake to be expected 6-7 months later on a preexisting fault across the seismicity gap.

View full abstract